system panic

system panic

Post by A.BENDAA » Sun, 21 Oct 2001 17:49:21



I have an AIX system that went in a panic
3 times and after the boot a do a diag
that says all hardware is OK
how can I check if there s a software problem

thank you


 
 
 

system panic

Post by Christer Pal » Sun, 21 Oct 2001 20:23:21



> I have an AIX system that went in a panic
> 3 times and after the boot a do a diag
> that says all hardware is OK
> how can I check if there s a software problem

Check the error log (errpt). If there's a Data Storage Interrupt
(DSI_PROC) or Program Storage Interrupt (PSI_PROC?) then you have a
software problem (as the errpt output will also tell you). In that case
you would also likely have a system dump (sysdumpdev -L) that would help
you diagnose the problem. Some basic dump analysis is done automatically
when the system comes back up again and the result is put in the error
log. Look for the "symtom string" in the error log entry and use that to
search IBM's support site.

--
Christer Palm

 
 
 

system panic

Post by Christer Pal » Wed, 24 Oct 2001 09:25:56




> > Check the error log (errpt). If there's a Data Storage Interrupt
> > (DSI_PROC) or Program Storage Interrupt (PSI_PROC?) then you have a
> > software problem (as the errpt output will also tell you). In that case
> > you would also likely have a system dump (sysdumpdev -L) that would help
> > you diagnose the problem. Some basic dump analysis is done automatically
> > when the system comes back up again and the result is put in the error
> > log. Look for the "symtom string" in the error log entry and use that to
> > search IBM's support site.

> Correction.  It's "Program Interrupt," not "Program Storage Interrupt."

Hmm... Thanks Nicholas.
Actually I was thinking about Instruction Storage Interrupt (ISI_PROC),
but I was obviously mixing it up with Program Interrupt (PROGRAM_INT)
which is something you should look for as well.
Additionally there's a few other things that also indicate software
problems that you should look out for, such as KERNEL_PANIC.

--
Christer Palm

 
 
 

system panic

Post by Christer Pal » Wed, 24 Oct 2001 16:38:42



> I forget how DSI and PI map to the LED numbers at crash time.
> One of them is 300, the other 700.  Do you remember which goes
> along with which?

IIRC, 102 700 is a DSI (which, in my experience, is what you see in like
90% of all cases).

--
Christer Palm

 
 
 

1. This is disconcerting: System panic: panic (cpu 0): Processor Machine Check

Every once in a while I will get the above message and the machine will
halt.

My system:

Ethernet and a 20GB IDE hard disk and IDE Sony CD-ROM.

Looking for more information on this, I have just been finding the extremely
helpful, "contact your service provider."
Taking a guess, this seems reminiscent of the old parity-error messages I
would get on my 386 with less-than-perfect memory.

I am just hoping that there could be a non-hardware related problem causing
this.
Or is my being terminally poor still manifesting itself by the creative
failure of the hardware I can afford?

- Steve Young

2. Shared Objects entry point

3. System panics on new ASUS TP4 pentium 90 system

4. Screen problems on a Dell 486p/33

5. Help Re: System Panics

6. 4D Window Manager?

7. System Panic Messages

8. PPP over UDP

9. System panic on E3500, Solaris 7

10. system panic

11. memory corruption - SYSTEM PANIC

12. System panics with parity error after upgrade

13. Linux/Dos file-system Panic